Retrieval of Land Surface Temperature from Remote Sensing Thermal Images

1
Retrieval of Land Surface Temperature from
Remote Sensing Thermal Images

• Peng Gong and Ruiliang Pu
• CAMFER
• Department of ESPM
• University of California, Berkeley, USA

145 Mulford Hall, Berkeley, CA 94720-31104 Phone
[phone] Fax [phone] E-mail
rpu_at_nature.berkeley.edu
2
Contents

• ? Introduction
• ? Objectives
• ? Four Thermal Sensors
• ? LST Retrieved from TABI
• ? LST Retrieved from ASTER
• ? LST Retrieved from AVHRR
• ? LST Retrieved from TM(ETM)
• ? Concluded Notes

3
Introduction

• ? Briefly recall the LST retrieval project from
  PASCO, Japan
• ? Importance of LST retrieved from remote sensing
  thermal images for studies of physical, chemical
  and biological processes of the Earth
• ? Two categories of LST retrieval methods
• Mono-window algorithm (MWA)
• Split-window algorithm (SWA)

4
Objectives

• ? Test and develop both mono-window algorithms
  and split-window algorithms for retrieving LST
  from thermal remote sensing images
• ? Develop softwares for calculating LST from 4
  thermal sensors data TABI-320, ASTER Ch13 14,
  AVHRR Ch4 5, and TM6 or ETM6

5
Four Thermal sensors

• ? TABI-320
• Thermal Airborne Broadband Imager, manufactured
  by ITRES
• Covering thermal range 8-12 mm, 320 pixels per
  line, brightness temperature data provided in Deg
  C X 100
• Altitude up to 3048 m

• ? ASTER
• Terra/ASTER Ch13(10.25-10.95 mm),
  Ch14(10.95-11.65 mm)
• Spatial resolution 90 m
• Be provided with scaled thermal radiance at sensor

6
Four Thermal sensors

• ? AVHRR
• NOAA11,14/AVHRR Ch4(10.5-11.3 mm),Ch5(11.5-12.5
  mm)
• Scaled thermal radiance at sensor
• Nominal spatial resolution 1.1 km at nadir

• ? TM(ETM)
• All Landsat 5 TM6 and Landsat 7 ETM6 after July,
  2001
• Scaled thermal radiance (DN) with a wavelength
  range 10.45-12.50 mm
• Spatial resolution 120 m for TM6 and 60 m for
  ETM6

7
LST retrieved from TABI

• ? Flowchart for LST retrieval
• ? Mono-window algorithm (MWA)
• ? Application cases
• ? Requirements of current version

8
LST retrieved from TABI
A flowchart for LST retrieved from TABI
9
LST retrieved from TABI
Mono-window algorithm for LST retrieval
10
LST retrieved from TABI
Transmittance simulated from water vapor with
MODTRAN4
11
LST retrieved from TABI
Emissivity assigned for 7 types of land cover
For different combinations of vegetation, soil
and water e0.965-0.985
Determination of Ta calculated from T0

• For the 4 standard atmospheric profiles
• 1976 USA Ta7.711000.96447T0
• Tropical Ta5.990010.97240T0
• Mid-Lat-N-Summer Ta4.732490.97819T0
• Mid-Lat-N-Winter Ta5.345390.97537T0.

12
LST retrieved from TABI
Application case Castle, Osaka, Japan
TABI BT image (oC X 100), 09/1/03, 750 am,
Min/Max/Average19.6/37.5/25.4oC
Retrieved LST image (oC X 100) 09/1/03, 750 am,
Min/Max/Average20.2/38.0/27.8oC
13
LST retrieved from TABI
Application case Center, Osaka, Japan
TABI BT image (oC X 100) Min/Max/Average13.0/31.2
/26.1oC
Retrieved LST image (oC X 100) Min/Max/Average13.
5/35.2/28.6oC
14
LST retrieved from TABI
Requirements of current version

• ? Altitude around 1 kilometer
• ? Total atmospheric water vapor (g/cm2), measured
  from ground, for the 4 profiles 0.2-2.8 for 1976
  USA, 0.2-5.4 for Tropical, 0.2-4.4 for Mid-lat-N
  summer, and 0.2-1.2 for Mid-lat-N winter
• ? Input TABI brightness temperature in oC X 100
• ? Use a fixed value of emissivity over the entire
  scene of TABI image

15
LST retrieved from ASTER

• ? Flowchart for LST retrieval
• ? Split-window algorithm (SWA)
• ? Application case
• ? Requirements of current version

16
LST retrieved from ASTER
A flowchart for LST retrieved from ASTER
17
LST retrieved from ASTER
Split-window algorithm for LST retrieval
Where,
18
LST retrieved from ASTER
Calculation of BT for ASTER Ch13 Ch14
Spectral radiance at sensor for Level 1B data
Brightness temperatures for ASTER 13 14
Where,
19
LST retrieved from ASTER
Transmittance simulated from wv with MODTRAN4 (I)
20
LST retrieved from ASTER
Transmittance simulated from wv with MODTRAN4 (II)
21
LST retrieved from ASTER
Emissivity assigned for 7 types of land cover
For different combinations of vegetation, soil
and water e0.954-0.988
22
LST retrieved from ASTER
Application case Tokyo, Japan
ASTER Ch 13 scaled radiance, September 05, 2003,
at 1035 am
23
LST retrieved from ASTER
Application case Tokyo, Japan
ASTER Ch14 scaled radiance, September 05, 2003,
at 1035 am
24
LST retrieved from ASTER
Application case Tokyo, Japan
Retrieved LST image (oC X 100) Min/Max/Average20.
0/74.5/38.4oC
25
LST retrieved from ASTER
Requirements of current version

• ? Total atmospheric water vapor (g/cm2), measured
  from ground, for the 4 profiles 0.2-2.8 for 1976
  USA, 0.2-6.6 for Tropical, 0.2-5.2 for Mid-lat-N
  summer, and 0.2-1.2 for Mid-lat-N winter
• ? Input two ASTER thermal channels in digital
  number (scaled radiance)
• ? Use a fixed value of emissivity over the entire
  scene of ASTER image

26
LST retrieved from AVHRR

• ? Flowchart for LST retrieval
• ? Split-window algorithm (SWA)
• ? Application case
• ? Requirements of current version

27
LST retrieved from AVHRR
A flowchart for LST retrieved from AVHRR
28
LST retrieved from AVHRR
Split-window algorithm for LST retrieval
(Based on Qin et al. (2001, JGR,
106(D19)22655-22670)
Where,
29
LST retrieved from AVHRR
Transmittance calculated from water vapor and ZAV
(q)
30
LST retrieved from AVHRR
Emissivity assigned for 7 types of land cover
For different combinations of vegetation, soil
and water
31
LST retrieved from AVHRR
Application case San Francisco, USA
AVHRR Ch5 BT image (K), September 02, 1999, at
300 pm
AVHRR Ch4 BT image (K), September 02, 1999, at
300 pm
32
LST retrieved from AVHRR
Application case San Francisco, USA
Retrieved LST image (oC X 100) Min/Max/Average7.6
/51.6/25.4oC
33
LST retrieved from AVHRR
Requirements of current version

• ? The algorithm with provided parameters values,
  either simulated or assigned, can be employed in
  the low- to middle-latitude regions of the Earth.
  
• ? Total atmospheric water vapor (g/cm2), measured
  from ground, for the 2 profiles, summer and
  winter is 0.2-6.4
• ? Input scaled brightness temperatures of both
  AVHRR channels (Ch 4 Ch5)
• ? Use a fixed value of emissivity over the entire
  scene of AVHRR image

34
LST retrieved from TM(ETM)

• ? Flowchart for LST retrieval
• ? Mono-window algorithm (MWA)
• ? Application case
• ? Requirements of current version

35
LST retrieved from TM(ETM)
A flowchart for LST retrieved from TM(ETM)
TM(ETM)6 scaled radiance at sensor
TM (ETM) 6 brightness temperature (BT) image
LST retrieval image
36
LST retrieved from TM(ETM)
Mono-window algorithm for LST retrieval
(Based on Qin et al. (2001, IJRS,
22(18)3719-3746)
Where,
37
LST retrieved from TM(ETM)
Calculation of BT for TM (ETM)6
Spectral radiance at sensor
Brightness temperatures for TM6 or ETM6
Note Using level 1G image for ETM6 after July,
2000
38
LST retrieved from TM(ETM)
Transmittance simulated from wv with LOWTRAN7
For transmittance for summer and winter
profiles, see Table 5 (p3733) by Qin et al.
(2001, IJRS) for linear calculation equations of
transmittance from water vapor.
Emissivity assigned for 7 types of land cover
For different combinations of vegetation, soil
and water e0.945-0.990
Determination of Ta calculated from T0
For the 4 standard atmospheric profiles 1976
USA Ta25.93960.88045T0 Tropical
Ta17.97690.91715T0 Mid-Lat-N-Summer Ta16.
01100.92621T0 Mid-Lat-N-Winter
Ta19.27040.91118T0.
39
LST retrieved from TM(ETM)
Application case Lake Tahoe, CA/NV, USA
ETM6 scaled radiance, October 25, 2001, at 1030
am
40
LST retrieved from TM(ETM)
Application case Lake Tahoe, CA/NV, USA
Retrieved LST image (oC X 100) Min/Max/Average1.1
/45.7/19.1oC
41
LST retrieved from TM(ETM)
Requirements of current version

• ? The algorithm with provided parameters values,
  either simulated or assigned, can be employed in
  most of the low- to middle-latitude regions of
  the Earth.
• ? Total atmospheric water vapor (g/cm2), measured
  from ground, for the 2 profiles, summer and
  winter is 0.4-3.0
• ? Input scaled radiance (DN) of TM6 or ETM6
• ? Use a fixed value of emissivity over the entire
  scene of TM6 or ETM6 image

42
Concluded Notes
? Both SWA and MWA for AVHRR and TM thermal
images, developed by Qin et al. (2001), have been
verified and validated properly. However,
? Both MWA and SWA for TABI and ASTER thermal
images, developed in this project, have not been
properly validated with ground truth.
? For current version of the MWA for TABI, the
altitude is assumed around 1 km. If changing the
altitude significantly, the estimation equations
of transmittance from water vapor and derivation
of Ta from T0 need to be re-simulated.
? Using a fixed emissivity over the entire scene
of thermal image may be problematic for
retrieving pixel-based LST.
43
Acknowledgments
? Thermal image samples of TABI and ASTER
provided by Ryo Michishita, PASCO Co., Japan.
? Both MWA and SWA algorithms for AVHRR and TM
developed by Qin et al. (2001, JGR and IJRS).
44
Thank you!